Examining Competitive Interactions Between Rasberry Crazy Ants
Transcript of Examining Competitive Interactions Between Rasberry Crazy Ants
RICE UNIVERSITY
Examining Competitive Interactions Between Rasberry Crazy Ants (Paratrechina sp. nr. pubens) and Red Imported Fire Ants (Solenopsis
invicta) Using Laboratory and Field Studies
by
Katherine Horn
A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE
Master of Arts
APPROVED, THESIS COMMITTEE:
Evan Siemann, Professor, Chair Ecology and Evolutionary Biology
(3-v a David Queller, Harry C. and OlgaK. Wiess Professor Ecology and Evolutionary Biology
y Djttrinam, Evolutional
Amy Djartnam, Assistant Professor Ecology and Evolutionary Biology
Micky Evmanks, Associate Professor Entomology, Texas A&M
HOUSTON, TEXAS AUGUST 2009
1
UMI Number: 1486923
All rights reserved
INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMT Dissertation Publishing
UMI 1486923 Copyright 2010 by ProQuest LLC.
All rights reserved. This edition of the work is protected against unauthorized copying under Title 17, United States Code.
ProQuest LLC 789 East Eisenhower Parkway
P.O. Box 1346 Ann Arbor, Ml 48106-1346
ABSTRACT
Examining Competitive Interactions Between Rasberry Crazy Ants {Paratrechina sp. nr. pubens) and Red Imported Fire Ants {Solenopsis
invicta) Using Laboratory and Field Studies
by
Katherine Horn
Studying normative species soon after their introduction is critical to
understanding their risk of becoming widely invasive and determining effective methods
of control. I conducted laboratory and field experiments on the Rasberry crazy ant
{Paratrechina sp. nr. pubens), which was introduced to Pasadena, TX in 2002 and has
since been spreading rapidly. These experiments focused on intraspecific aggression, as
well as individual and colony-level interactions between crazy ants and red imported fire
ants {Solenopsis invicta), which are dominant in the crazy ant's introduced range. Crazy
ants displayed no intraspecific aggression. In individual aggressive encounters with fire
ants, crazy ants had higher mortality than fire ants, but in colony-level clashes, crazy ants
had less mortality and better control of food resources. These findings suggest that crazy
ant abundance is key to their competitive success, and fire ants may provide biotic
resistance to crazy ants in some areas.
ii
ACKNOWLEDGEMENTS
I thank my advisor, Evan Siemann, for accepting me as a student in his lab at a
turning point in my graduate school experience. He has always respected my ideas and
provided me with supportive feedback and guidance, and he has created an environment
that has allowed me to thrive. My other committee members; Amy Dunham, Micky
Eubanks, and Dave Queller; have also provided me with valuable input throughout the
conception, execution, and writing of this thesis. I also thank the Wray-Todd family for
their generous support of ecology and evolutionary biology students at Rice.
My labmates Juli Carrillo, Christopher Gabler, and Maria Meza-Lopez have been
incredibly encouraging, especially as my defense grew closer. Juli in particular was an
amazing cheerleader, and I couldn't have asked for a more wonderful person to share
responsibilities and recipes with during my two years as a graduate student. I will miss
seeing her everyday. I also thank all of the undergraduates who have contributed to the
success of my research. Matt Barajas, Marilu Corona, Summer Jones, Matt King, Max
Quintanilla, and Erica Soltero braved the heat on ant collecting trips to Katy and censuses
at the Coastal Center, suffered through fire ant stings without complaint, and counted
hundreds of tiny ants. Erica Soltero in particular was extremely dedicated to the cause
and was a joy to be around everyday. Her enthusiasm for the research is a sign of the
success she will surely have in the future and was a great motivation to me when I was
dragging my feet. As I have said before, there are not enough Sonic sundaes in the world
to sufficiently thank her for her help.
I also want to express my gratitude to my family. My older sister Elizabeth
provided long-distance encouragement from China, and my younger sister Johanna
i i i
provided assistance in the field and much-needed distractions during the final few months
of the thesis process. My father, John, also managed to call and cheer me up at just the
right moments and to ask me science questions that reminded me why I started the thesis
in the first place. My stepfather Jim always had supportive words. And when my mother
Annette unknowingly let me raise my first ant colony in my bedroom on a stash of hidden
Halloween candy, she probably could not have guessed that one day I would be studying
ants full time. I appreciate her interest in my work and her warm and kind words that
urged me on so many times when I doubted myself.
My friends in Houston were also incredibly supportive. The Panchang family
provided me a home away from home, though sadly I was too busy to go there as often as
I would have liked. All of my friends were always genuinely curious about my ant
battles; I ended up talking about ants at many get-togethers, and yet they never seemed to
grow tired of it. Towards the end of the process, their endless supply of hugs, their
understanding when I had to turn down nearly all of their invitations, and their collective
chorus of "you can do it" gave me the boost I needed.
Last but not least, I acknowledge the tireless support of my fiance, Scott
Chamberlain. He was an invaluable resource on all ant-related questions, and he always
found time to help me, even during his own very busy field season. More importantly, he
was never short on kind words, and he brought me back from the edge of a breakdown on
many occasions. He also did everything in his power to minimize my burdens outside of
work during the last couple of months by doing laundry, planning meals, and keeping the
house clean. I truly could not have done this thesis without him, and I want him to know
that I appreciate all that he did to keep me happy, healthy, and focused on my thesis.
iv
TABLE OF CONTENTS
Chapter
1. Introduction 1
2. The Effect of Diet on Intra- and Interspecific Interactions of the Rasberry Crazy Ant (Paratrechina sp. nr. pubens) .8
3. Colony-Level Competition Between Rasberry Crazy Ants {Paratrechina sp. nr. pubens) and Red Imported Fire Ants (Solenopsis invicta) 36
v
LIST OF TABLES
2.1. Head lengths of fire ants and crazy ants and fire ant colony size distribution 31
3.1. Tests for each response variable for each comparison 61
3.2. Significant correlations between response variables ...62
vi
LIST OF FIGURES
2.1. The effect of fire ant worker size on the average number of fire ants fighting during each observation (out of a total of 5 fire ants) 32
2.2. The number of Rasberry crazy ants and fire ants engaged in fights 33
2.3. The number of dead Rasberry crazy ants and dead fire ants after 1 hour 34
2.4. The effect of sugar level in crazy ant diet on the average aggression score for each set of fire ants and crazy ants 35
3.1. Mass of dead ants at the end of the colony-level competition experiment 63
3.2. Average sugar water consumption by species, adjusted for average desiccation 64
3.3. Recruitment to mealworms and sugar water 65
3.4. Mealworm and sugar water recruitment times 66
3.5. Ants fighting and ants in the other nest box. 67
3.6. Changes in ant recruitment to field baits through time 68
3.7. Average ant recruitment to field baits by treatment and bait type from June 24 to July 14 69
3.8. Discovery time of field baits by bait type and species 70
vii
Chapter 1
Introduction
Invasive species impact ecosystems and the economy. In the United States alone,
control efforts directed at invasive species and environmental damage caused by invasive
species total $137 billion annually (Pimentel et al. 2000). Ants are among the worst
invaders and have caused ecological damage and economic costs worldwide (Lowe et al.
2004). Invasive ants often become highly abundant in their introduced ranges,
outnumbering native ants and reducing native ant diversity (Holway et al. 2002), as well
as affecting other invertebrates and vertebrates such as birds and lizards (Holway et al.
2002, Kenis et al. 2009). Preventing introductions is the most effective way to control
the spread of invasive species, yet non-native ants continue to be accidentally introduced
to new environments around the world. Therefore it is important to study the dynamics
of new introductions in order to assess the risk of widespread invasion and to determine
the most effective methods of control, as resources for controlling invasive species are
often limited.
Despite considerable research on invasive species, one area that needs further
exploration is how established invasive species influence the establishment and spread of
newly introduced species. In some cases, existing invasive species have been shown to
facilitate new invasions, which can lead to invasional meltdown (Simberloff and Von
Holle 1999, O'Dowd et al. 2003, Bourgeois et al. 2005). However, in cases where
established invasive species act as competitors or predators of newly introduced species,
the established species may provide biotic resistance to new invasions (Simberloff and
Von Holle 1999). Nevertheless, very few studies have looked at competition between
1
invasive species in their shared introduced range (but see Belote and Weltzin 2006 and
Griffen et al. 2008), and to my knowledge no studies have explicitly tested the negative
effects of long established species on new invaders. Therefore, I tested for biotic
resistance to a new introduction of a non-native species by an established invasive
species, using ants as a study system.
In 2002, the Rasberry crazy ant, Paratrechina sp. nr. pubens, was unintentionally
introduced to an industrial area along a ship channel in Pasadena, TX (Meyers 2008).
Since its introduction to Texas, populations of Rasberry crazy ants have spread by 20-
30m per month (Meyers 2008), and as of 2008 crazy ants had been reported in the
following Texas counties: Bexar, Brazoria, Fort Bend, Chambers, Galveston, Hardin,
Harris, Jefferson, Jim Hogg, Liberty, Montgomery, Orange, Walker and Wharton (Texas
A&M 2008). The high rate of range expansion, along with the presence of extremely
high densities of crazy ants in invaded areas, suggest that this exotic ant has the potential
to become widespread and invasive. Additionally, very few pest control products, none
of which are directly available to consumers, have been found to be effective in
controlling crazy ant invasions (Texas A&M 2008).
Little is known about this species, including its native range and precise
identification. Taxonomists have observed morphological and behavioral differences
between Texas populations of P. sp. nr. pubens and the previously described P. pubens,
yet the classification of P. sp. nr. pubens as P. pubens or an undescribed species remains
inconclusive (Meyers 2008). The biology of the crazy ant is largely unknown, but the
species is known to be polygynous and to nest ephemerally. Though crazy ant colonies
2
are known to produce winged males and females, nuptial flights have never been
observed in the native range (Texas A&M 2008).
Because so little is known about the ecology of this exotic ant, I conducted a
series of laboratory and field experiments designed to explore the nature of the
intraspecific interactions between Rasberry crazy ants, as well as the individual and
colony-level competitive ability of crazy ants in comparison to the dominant ant species
in its introduced range, the red imported fire ant, Solenopsis invicta, which is also an
introduced species. Together these experiments addressed the issues of unicoloniality,
the effect of diet on aggressive encounters, resource acquisition, and most importantly,
biotic resistance by fire ants to the crazy ant invasion. In Chapter 2,1 performed a series
of aggression assays between crazy ants and fire ant workers of different sizes, between
crazy ant workers from colonies receiving one of two protein sources and high or low
doses of sugar, and between fire ants and crazy ant workers receiving the different diets.
I found that large fire ants engaged in fewer fights than small fire ants, but fire ant size
affected neither fire ant nor crazy ant mortality. On average, crazy ants experienced
higher mortality than fire ants after antagonistic interactions. Diet did not induce
aggressive interactions between crazy ants, but sugar level did play an important role in
antagonistic interactions with fire ants, as crazy ants on a low sugar diet earned higher
aggression scores.
The experiments in Chapter 3 examined colony-level interactions between crazy
ants and fire ants both in the field and in the laboratory. In the laboratory experiment,
colonies of fire ants and crazy ants were standardized by worker number or by biomass
and used as either controls or in competing pairs that had access to a single set of protein
3
and sugar baits. For both colony sizes, and in both control and competition settings,
crazy ants recruited more quickly and in higher abundances to baits. Crazy ants also
controlled significantly more baits than fire ants did at the end of the experiment. These
results suggest that crazy ants, when evenly matched in colony size with fire ants, are
capable of both discovering baits first and excluding fire ants from them, a combination
of traits that has not been reported in native ants, but has been shown in invasive
Argentine ants (Davidson 1998). To further test these findings, I conducted a field
experiment where small crazy ant colonies were moved from a heavily invaded site to
another location 750m away that had not yet been invaded. I then monitored the ant
recruitment to sugar and protein baits for three weeks. Crazy ants discovered baits more
quickly, but fire ants controlled the majority of baits and recruited to both sugar and
protein in higher abundances than crazy ants. Crazy ant recruitment increased through
time in crazy ant addition locations, but fire ant recruitment decreased in both control and
crazy ant addition locations, suggesting that crazy ants are responding to, rather than
causing, drops in fire ant recruitment.
In sum, the results of Chapters 2 and 3 suggest that crazy ant abundance is likely
an important factor in competitive interactions with fire ants. On an individual level, fire
ants are more successful fighters than crazy ants, but as colony units, crazy ants are more
successful at resource retrieval and engage in more fights without experiencing correlated
mortality when crazy ant and fire ant colonies are of similar size. The field experiment
reinforced this hypothesis, as the small crazy ant colonies did not impact the recruitment
to baits by fire ants from established nests. These findings suggest that fire ants may
provide biotic resistance to crazy ants in areas that do not support high crazy ant
4
densities, a factor which may severely restrict the potential spread of crazy ant
populations. Putting my results in the context of observations from the field, I suggest
that the future of the crazy ant invasion in southeast Texas will likely consist of
coexistence with fire ants, likely in a patchy distribution of habitat partitioning caused by
environmental factors that affect crazy ant abundance. Further studies, such as large-
scale surveys of invaded sites and closer examinations of invasion fronts, would be useful
in providing additional merit to this hypothesis. Nevertheless, these studies provide a
crucial first exploration of the future of the crazy ant invasion, as well as a demonstration
of the importance of competition between long-established invasive species and newly
introduced species.
Literature Cited
Belote, R. T. and J. F. Weltzin. 2006. Interactions between two co-dominant, invasive
plants in the understory of a temperature deciduous forest. Biological Invasions
8:1629-1641.
Bouregois, K., C. M. Suehs, E. Vidal, and F. Medail. 2005. Invasional meltdown
potential: facilitation between introduced plants and mammals on French
Mediterranean islands. Ecoscience 12:248-156.
Davidson, D. W. 1998. Resource discovery versus resource domination in ants: a
functional mechanism for breaking the trade-off. Ecological Entomology 23:484-
490.
5
Griffen, B. D., T. Guy, and J. C. Buck. 2008. Inhibition between invasives: a newly
introduced predator moderates the impacts of a previously established predator.
Journal of Animal Ecology 77:32-40.
Holway, D., L. Lach, A. Suarez, N. Tsutsui, and T. Case. 2002. The causes and
consequences of ant invasions. Annual Review of Ecology, Evolution, &
Systematics 33:181-233.
Kenis, M, M. Auger-Rozenberg, A. Roques, L. Timms, C. Pere, M. Cock, J. Settele, S.
Augustin, and C. Lopez-Vaamonde. 2009. Ecological effects of invasive alien
insects. Biological Invasions 11:21-45.
Lowe, S., M. Browne, S. Boudjelas, and M. De Poorter. 2004. 100 of the world's worst
invasive species: a selection from the Global Invasive Species Database. Invasive
Species Specialist Group, Gland, Switzerland.
Meyers, J. M. 2008. Identification, distribution and control of an invasive pest ant,
Paratrechina sp. (Hymenoptera, Formicidae), in Texas. PhD Thesis. Texas A&M
University, College Station, TX.
O'Dowd, D. J., P. T. Green, and P. S. Lake. 2003. Invasional 'meltdown' on an oceanic
island. Ecology Letters 6:812-817.
Pimentel, D., L. Lach, R. Zuniga, and D. Morrison. 2000. Environmental and economic
costs of nonindigenous species in the United States. Bioscience 50:53-65.
Simberloff, D. and B. Von Holle. 1999. Positive interactions of nonindigenous species:
invasional meltdown? Biological Invasions 1:21-32.
6
Texas A&M, Center for Urban & Structural Entomology. 2009. Rasberry Crazy Ant,
Paratrechina sp. nearpubens. http://urbanentomology.tamu.edu/ants/exotic
_tx.cfm (October 25,2009).
Chapter 2
The Effect of Diet on Intra- and Interspecific Interactions of the Rasberry Crazy Ant {Paratrechina sp. nr. pubens)
Abstract
Biotic interactions are often important in the establishment and spread of invasive
species. In particular, competition with species in the introduced range of an exotic
species can be significant in preventing an exotic species from colonizing areas of
suitable habitat (Walters and Mackay 2005, Menke et al. 2007). Here we examined
interference competition in the Rasberry crazy ant, Paratrechina sp. nr. pubens, which
was recently introduced to the Gulf Coast of Texas, by studying intraspecific interactions
between nests of Rasberry crazy ants and interspecific interactions with the red imported
fire ant, Solenopsis invicta. Specifically, we examined individual-level interactions in
three sets of ants: 1) between crazy ant workers and either small (minor) or large (major)
fire ant workers, 2) between crazy ant workers from different nests receiving one of two
protein sources and either high or low doses of sugar, and 3) between fire ants and crazy
ant workers receiving the different diets. We found that large fire ants fought less often
with crazy ants than did small fire ants, but fire ant size affected neither fire ant nor crazy
ant mortality. Overall, crazy ants experienced higher mortality than fire ants after
antagonistic encounters. Differences in diet did not alter interactions between crazy ant
workers from different nests, but sugar level did play an important role in antagonistic
interactions with fire ants, as crazy ants on a low sugar diet earned higher aggression
scores. Our findings suggest that fire ants may outcompete crazy ants on an individual
level, providing some biotic resistance to crazy ant range expansion. However, fire ants
8
may overcome by individual crazy ants that have a restricted sugar intake, which may
occur when crazy ants are excluded from resources by fire ants.
Introduction
Biotic interactions between introduced species and native or pre-established
exotic species are important influences on the success and spread of invasive species
(Mack et al. 2000, Mitchell and Power 2003, Ness and Bronstein 2004, Torchin and
Mitchell 2004). Competition from established species in the introduced range, like
between ant species for food or territory, can sometimes serve as biotic resistance to
invasive species (Walters and Mackay 2005, Menke et al. 2007). Competitive
interactions can be exploitative, where individuals compete indirectly via their effects on
shared resources, or interference, where individuals directly clash through antagonistic
behaviors (Ricklefs 2001). In ants, interference competition is common and can cause
death or injury of workers and loss of access to food or territory (Savolainen and
Vepsalainen 1988, HoTldobler and Wilson 1990, Andersen and Patel 1994). Therefore,
interference competition can be an important factor in the spread and success of an exotic
ant pest, such as the recently introduced Rasberry crazy ant, Paratrechina sp. nr. pubens.
In this paper, I examine direct intraspecific competitive interactions between crazy ant
nests and direct interspecific competitive interactions between crazy ants and an
extremely common invasive ant in its introduced range, Solenopsis invicta, the red
imported fire ant. Intraspecific and interspecific interactions of Rasberry crazy ants have
never been examined, yet the ant is a pest species that is increasing its introduced range
southeast of Houston, TX. To my knowledge, no studies have examined competition
9
between long established invasive species and newly introduced species for evidence of
biotic resistance, yet the invasive fire ants may provide biotic resistance to the spread of
Rasberry crazy ants. Therefore it is important to investigate antagonistic encounters both
between Rasberry crazy ants nests and with fire ant workers.
Because nestmate recognition, territoriality, and activity level are influenced by
diet (Davidson 1997, Howard and Blomquist 2005, Buczkowski and Silverman 2006,
Grover et al. 2007), the diet of an ant can influence aggressive interactions, whether
between species or colonies of a single species. The cuticular hydrocarbons of prey items
can be transferred to ants that come into contact with it, disrupting the ant's hydrocarbon
profile and affecting nestmate recognition (Buczkowski et al. 2005). Differences in prey
type have been shown to induce aggression between nestmates (Silverman and Liang
2001, Corin et al. 2007, Sorvari et al. 2008) and to affect antagonistic interactions
between nests or colonies of a given species (Liang et al. 2001, Buczkowski and
Silverman 2006, Grover et al. 2007), though the effect of diet on interspecific
antagonistic interactions has not been extensively studied (but see Alloway et al. 1991),
and the effect of diet on competitive interactions between two invasive ant species has
never been studied. Additionally, it has been proposed that invasive ants consume higher
amounts of carbohydrate resources than native ants, which may increase the tempo of
their worker activity, contributing to their high levels of abundance and domination of
resources and territory (Holway et al. 2002). The amount of sugar in a colony's diet has
been shown to be an important variable in activity level and antagonistic behavior in both
Formica aquilonia and invasive Argentine ants, Linepithema humile (Grover et al. 2007,
Sorvari et al. 2008). Because of the importance of diet in interactions between ants, we
10
examined the effects of both sugar level and protein type on intra- and interspecific
competition in Rasberry crazy ants.
Paratrechina sp. nr. pubens (Hymenoptera, Formicidae, Formicinae) was first
discovered in an industrial area along the ship channel in Pasadena, TX in 2002 (Meyers
2008). Little is known about this ant species, including its exact species identification
and its native range. Ant taxonomists have noted differences between Texas populations
of Rasberry crazy ants and the previously described P. pubens, yet morphological and
behavioral differences between these ants have proven to be inconclusive in determining
if P. sp. nr. pubens is P. pubens or an undescribed species (Meyers 2008). Since its
introduction to Texas, the range of Rasberry crazy ants has increased by 20-30m per
month (Meyers 2008). Media reports on the ant (named after the first pest control agent
to discover the invasion, Tom Rasberry) focus on the tendency of Rasberry crazy ants to
nest in electronics and cause short circuits. Though these effects of the crazy ants are
likely overstated, it is known that Rasberry crazy ants are often found in extremely high
densities in invaded areas (Meyers 2008). Like another well-known invader, the
Argentine ant (Holway et al. 1998), crazy ant populations appear to be unicolonial;
colony boundaries seem to be nonexistent as ants move freely between nests (Holldobler
and Wilson 1990). Crazy ants are often found in areas that would be suitable habitat for
red imported fire ants (personal observation, K. Horn), such as woods and open areas,
suggesting that the two species may often come into contact and compete for resources.
Red imported fire ants (Hymenoptera, Formicidae, Myrmicinae) are themselves
one of the world's top 100 worst invasive species (Lowe et al. 2004). Introduced to the
United States via the port city of Mobile, AL in the 1930's (Buren et al. 1974), fire ants
11
have since spread throughout more than 106 million hectares of the southeastern US, the
Midwest, and California, where they are the dominant ant species in disturbed habitats
(Williams 1994). Due to their harmful effects on humans, agriculture, and ecosystems,
fire ants cost nearly $1 billion per year in economic losses and control efforts (Pimentel
et al. 2000). Some news reports have suggested that crazy ants attack, eat, and displace
fire ants, yet none of these claims have been tested. Due to the widespread invasion of
fire ants and the locally abundant populations of Rasberry crazy ants, interactions
between these two species may be very important in affecting the spread of crazy ant
populations.
To examine the intra- and interspecific interactions of the Rasberry crazy ants, the
following sets of ant pairings were observed: 1) crazy ants and either small (minor) or
large (major) fire ants workers, 2) crazy ants from nests in the same supercolony which
had been isolated and fed one of two protein sources and either high or low doses of
sugar water, and 3) crazy ants which had been on the different diets described above and
fire ants. Two sizes of fire ants were used because both minor and major fire ants forage
outside of the nest where they can interact with crazy ant foragers, and the different sizes
of fire ants could have different fighting tactics or abilities which may affect their
likelihood to engage in fights and/or their ability to kill ant opponents. All interactions
were examined using aggression assays with 5 ants from each species or treatment in a
Petri dish, which have been shown to be highly consistent and correlate well with full
colony introductions (Roulston et al. 2003). Together, these three sets of aggression
assays made it possible to address the following questions: 1) Are crazy ants or fire ants
more successful in fights? 2) Is fighting success affected by fire ant size? 3) Do
12
differences in diet affect interactions between workers of different crazy ant colonies?
and 4) Can the diet of crazy ants affect antagonistic interactions with fire ants? Question
three examines whether crazy ant unicoloniality can be affected by diet, while questions
one, two, and four focus on whether fire ants may provide biotic resistance to further
crazy ant colonization and spread. Together, the answers to these questions could have
important implications for the future of the crazy ant introduction.
Methods
All colonies of Rasberry crazy ants used in aggression assays were collected in
Pearland, TX (29.55°N, 95.28°W) on May 24 and May 31,2008. Though 24 distinct
nests were collected, because crazy ants at this site display no aggression between nests,
it is likely that all nests are part of a large supercolony (Holldobler and Wilson 1990).
Fire ant colonies were collected from Katy, TX (29.93°N, 95.94°W) in early May 2008.
Colonies of both species were polygynous. In order to separate ants from the nesting
material with which they were collected, we flooded colonies and then transferred all ants
to 24cm x 11cm plastic nest boxes. Nest boxes were ringed with a thin layer of
Tanglefoot (Tanglefoot, Grand Rapids, Michigan, USA) near the top of the inside walls
in order to prevent escapes.
The initial diet administered to both crazy ant and fire ant colonies consisted of
freezer-killed mealworms and sugar water. This diet was maintained until July 18,2008,
when all food was removed from the crazy ant nest boxes and diet manipulations began.
Each of the 24 crazy ant colonies was assigned to one of four treatments: cricket/high
sugar, cricket/low sugar, wax worm/high sugar, or wax worm/low sugar, such that there
13
were six colonies in each treatment. The low concentration sugar water consisted of 5 ml
of sugar in 120 ml of water (4.2% sugar by volume); the high concentration sugar water
consisted of 20 ml of sugar in 120 ml of water (16.7% sugar by volume). Each colony
was given half of a freezer-killed cricket (Orthoptera, Gryllidae) or half of a freezer-
killed wax worm and 7.4 ml of sugar water of the appropriate concentration every other
day. Wax worms (Lepidoptera, Pyralidae) were supplied by Armstrong's Cricket Farm
(West Monroe, LA) and crickets were purchased at a pet supply store. These two bait
items have different nutritional value; wax worms are less than 23% protein by mass,
while crickets are more than 50% protein by mass (Finke 2002, Wang et al. 2004). While
crazy ants were receiving experimental diets, fire ant colonies were starved.
The following aggression assays were performed: 5 crazy ant workers in a Petri
dish with either 5 minor or 5 major fire ant workers, 5 crazy ant workers from one diet
treatment in a Petri dish with 5 crazy ants from a colony receiving a different diet
treatment, and 5 minor fire ants with 5 crazy ants on an experimental diet. For each
aggression assay, ants were placed in a 9-cm Petri dish with Fluon (polytetrafluoro-
ethylene)-coated sides that prevented ant escape during these periods of observation. The
Suarez scale (Suarez et al. 1999) was then used to score the behavior of pairs of
interacting ants every minute for either 5 or 10 minutes, depending on the species
combination: 0 -no interaction (i.e., ants ignored each other), 1 - touching with antennae,
2 - avoidance, 3 - aggression (such as biting antennae or legs), and 4 - fighting (both ants
engaged). The average aggression score for each pairing was then calculated using all
scores from the duration of observation, including the zeroes representing ants that were
not interacting. The number of ants from each treatment or colony engaged in fights was
14
also recorded each minute. In all combinations, the first observation was made five
seconds after all ants were released into the Petri dish. At the end of the 5- or 10-minute
observation period, the ants were then left in the Petri dish for 1 or 2 hours, and a
mortality score was recorded after each hour. All statistical analyses were performed
using JMP 7.0.2 (SAS Institute Inc., 2007). Each of the aggression assays is explained in
further detail below.
Crazy ants vs. small and large fire ants
Aggression assays between crazy ant workers and fire ant workers of different
sizes were conducted on July 16,2008. Each of 12 fire ant colonies was used twice as a
source of workers, the first time with 5 small workers paired up with 5 crazy ant workers,
and the second time with 5 large workers paired up 5 crazy ant workers from a different
colony, creating a total of 24 aggression trails. Aggression and fighting data were
recorded for 10 minutes. Mortality was recorded after 10 minutes, and for 10 pairings,
also after 1 hour.
In order to quantify the size difference between crazy ants and minor and major
fire ant workers, the head length (from the front of the clypeus to the posterior margin of
the head) of 10 ants from each of the 3 groups was measured. Head length is considered
the most reliable predictor of body size across ant species (Kaspari and Weiser 1999).
The average and standard error of the length was calculated for each group, and the head
lengths were tested for a significant effect of ant type using a one-way ANOVA,
followed by a post hoc Tukey's test to discern between the 3 types. The head lengths of
each type of ant (crazy ant, small fire, and large fire ant) were all significantly different
from one another (F2.29 = 262.87, p = <0.0001, Table 2.1). In order to estimate the size
15
distribution of fire ant workers, between 21 and 36 fire ant workers from each of 10
colonies were classified as small, medium, or large. These classifications were based on
visual estimates based on the range of worker sizes, rather than specific measurement
ranges. The average percentage and standard deviation of workers in each category was
calculated (Table 2.1).
Overall average aggression score, average number of fire ants fighting, and
average number of crazy ants fighting all were analyzed using a one-way ANOVA to
determine the effect of fire ant size on the response variables. Average mortality after 10
minutes was minimal and often zero, thus it was not analyzed. Mortality for both crazy
ants and fire ants after 1 hour was non-normal and could not be transformed to achieve
normality and therefore was analyzed for an effect of fire ant size using a Wilcoxon
signed rank test. To determine if the average number of ants fighting differed by species,
the difference between species in number of ants fighting was taken for each aggression
assay and the resulting values were tested for significant deviation from zero using a
Wilcoxon signed rank test. To test for a species difference in mortality after 1 hour, the
differences were tested for deviation from zero using a 1-group, 2-tailed t-test. The
differences for both response variables were also tested for a fire ant size effect using
Wilcoxon signed rank tests. Additionally, average crazy ant and fire ant mortality after 1
hour was compared to survival of control Petri dishes containing either 5 crazy ants or 5
fire ants using Wilcoxon signed rank tests on the fire ant mortality and the square root of
crazy ant mortality after 1 hour.
Crazy ants receiving different diets
16
Intraspecific aggression trials were conducted on July 29,2008. A total of 24
pairings of colonies were used. Some source colonies were used more than once, but all
pairings were unique. There were 24 total pairings: 8 replicates of pairings with the same
sugar level but different protein type, 6 replicates of pairings with different sugar levels
but the same protein type, 6 replicates of pairings differing in sugar level and protein
type, and 4 replicates of pairings of the same sugar level and protein type. In order to
distinguish between the two treatments of crazy ants in aggression trials, workers were
coated with either pink or green fluorescent powder (Day Glo Color Corp. Cleveland,
Ohio, USA) using a small paintbrush in an intermediate holding container a few minutes
before they were added to the Petri dish. After the ants had ceased grooming and
resumed moving about the container, they were placed into the Petri dish. Aggression
and fighting data were recorded for 5 minutes, and the number of dead ants of each color
was recorded at 5 minutes, 1 hour, and 2 hours. As a control, the average mortality of un-
powdered crazy ants kept in a Petri dish for 2 hours was tested.
Because data were non-normal, even with transformations, a Wilcoxon signed
rank test was used to determine if peak interaction score varied significantly when colony
pairs differed by protein type, sugar level, both, or neither. The same test was also used
to determine if diet differences affected mortality after 1 or 2 hours. Additionally, a
Wilcoxon signed rank test was used to compare 1 and 2 hour mortality of all pairings of
different colonies and powdered controls of 10 workers from the same colony and
unmanipulated controls of 5 workers from the same colony.
Crazy ants receiving different diets vs. fire ants
17
In order to test the effect of crazy ant diet on interactions with fire ants,
aggression assays were again performed on July 31,2008, using 5 crazy ant workers and
5 small fire ant workers. Each crazy ant colony was used only once. Two fire ant
colonies were used twice, but each colony pair was unique. There were 6 aggression
assays performed for each of the 4 treatments for a total of 24 colony pairings.
Aggression and fighting data were recorded for 5 minutes, and mortality was recorded at
5 minutes, 1 hour, and 2 hours.
To determine the effect of diet on aggression score, fighting, and mortality of the
two species, 2-way ANOVA's were performed with protein type, sugar level, and the
interaction of protein type and sugar level as factors. The following response variables
had a normal distribution without transformation: the average interaction score across all
time periods, the average number of crazy ants in engaged in fights, the average number
of fire ants involved in fights, and crazy ant mortality after 2 hours. The mortality of fire
ants after 2 hours was square root transformed prior to testing for a diet effect with an
ANOVA. Mortality after 2 hours was used because it had higher values and a more
normal distribution than 1-hour mortality, and in only 2 cases were all 5 of the ants from
a colony in a Petri dish dead (both occasions were fire ants). In order to test for a
difference between crazy ants and fire ants in average number of workers fighting and
mortality after 2 hours, the fire ant values were subtracted from the crazy ant values for
each response variable. A Wilcoxon signed rank test was used to test if the mean of the
difference in numbers of workers fighting was significantly different from zero, and an
ANOVA was used to determine if the mean of the difference in mortality after 2 hours
was affected by sugar level, protein type, or an interaction of the two variables. Also, a
18
Wilcoxon signed rank test was used to see if protein or sugar level significantly affected
the difference between the average number of crazy ant and fire ant workers engaged in
fights.
Results
Crazy ants vs. minor and major fire ants
There was no effect of fire ant worker size category on overall average aggression
score (Fi;22= 1-20, p = 0.285). The average aggression score of containers with large fire
ants was 1.00 ± 0.11, and the average aggression score of dishes with small fire ants was
1.18 ± 0.11. Though these scores were low due to the number of ants ignoring the other
species (receiving a score of 0), nearly all of the non-zero interactions were scored with
3's or 4's, indicating a high level of aggression during encounters. Though there was no
effect of fire ant size category on average number of crazy ants fighting (Fi,22 = 2.91, p =
0.102), there was a significant effect of fire ant size on the average number of fire ants
fighting (Fi,22 = 4.76, p = 0.040), with small fire ants having an average of 51.6% more
workers fighting than large fire ants fighting during a given observation period (Figure
2.1). Fire ant size category had no effect on either crazy ant mortality (Zi = 0.95, p =
0.329) or fire ant mortality (Zi = 1.72, p = 0.190) after 1 hour.
Comparing the response variables of the two species, there was a trend for crazy
ants to have a higher average number of workers engaged in fights (Z23 = 12.00, p =
0.063, Figure 2.2), and crazy ants had higher mortality than fire ants after 1 hour fa =
2.4140, p = 0.039, Figure 2.3). Fire ant size did not have an effect on the difference
between crazy ants and fire ants for average number of ants fighting (Z\ = -1.23, p =
19
0.208) or mortality after 1 hour (Zi = -0.43, p = 0.670). Crazy ant mortality when
fighting fire ants was significantly higher than that of control crazy ants (Z\ = -2.65, p =
0.008). Controls averaged 0.2 ± 0.2 dead crazy ants after 1 hour, whereas crazy ants
fighting fire ants had an average mortality of 2.7 ± 0.45 ants after 1 hour. Likewise,
fighting fire ants had significantly higher mortality after 1 hour than controls (Zi = -2.01,
p = 0.037).
Crazy ants receiving different diets
There was no fighting between crazy ants in any of the aggression assays; the
highest aggression score was a 2, which occurred only 5 times during the 144 observation
periods. Peak interaction score therefore was not significantly affected by diet treatments
(Z3 = 1.57, p = 0.667). There was no mortality for any of the ants after 5 minutes.
Neither mortality after one hour (Z3 = 2.95, p = 0.340) nor mortality after two hours (Z3 =
2.42, p = 0.491) were significantly affected by diet treatments. The overall average
mortality after two hours was 1.09 ± 0.21 out of 5 ants. This was not significantly
different from the average two-hour mortality of the five sets of unpowdered ants in a
Petri dish (Zi = 0.15, p = 0.694). Additionally, the 4 pairs that received the same diet did
not have significantly different mortality than the 5 unpowdered controls (Z\ = 0.00, p =
1.000).
Crazy ants receiving different diets vs. fire ants
Sugar level significantly affected the average interaction score (Fi,2o = 4.97, p =
0.037), though protein type had no effect (Fi^o = 0.75, p = 0.398) and the interaction of
the two factors was not significant (Fi^o = 1.16, p = 0.294). Ants receiving a lower sugar
diet had higher aggression scores when interacting with fire ants than did high sugar
20
crazy ants (Figure 2.4). There was no effect of sugar level on the on the number of crazy
ants (Fi>20 = 3.16, p = 0.091) or fire ants (Fi)20 = 3.28, p = 0.085) engaged in fights.
Protein type and the interaction of the two resources were not significant factors in either
the number of crazy ants fighting (F13 = 0.10, p = 0.757 for protein; Fi^o = 1.27, p =
0.2730 for the interaction) or the number of fire ants fighting (Fi o = 0.22 and p = 0.644
for protein, Fi o = 0.95 and p = 0.341 for the interaction). After 2 hours, there was no
effect of sugar level on crazy ant mortality (Fi;2o = 3.19, p = 0.090). Neither protein type
(Fi,2o = 0.09, p = 0.769) nor the interaction of sugar and protein (Fi^o = 0.09, p = 0.769)
significantly affected crazy ant mortality, and none of the response variables affected fire
ant mortality (Fi o = 0.21, p = 0.651 for sugar level; Fi o = 2.84, p = 0.108 for protein
type; Fi,2o = 0.59, p = 0.453 for the interaction).
Overall, there was no difference between species in average number of workers
fighting (Z23 = 9.00, p = 0.289) or mortality (t23 = -1.14, p = 0.268). There was no effect
of sugar level (Fi,2o = 2.49, p = 0.130), protein type (F u o = 0.90, p = 0.354), or the
interaction of diet components (Fi,2o = 0.54, p = 0.470) on the difference between crazy
ants and fire ants in mortality. There was also no effect of protein or sugar level (both Zi
= -0.73, p = 0.444) on the difference in average number of fighting ants of each species.
Discussion
The results of the trials between crazy ants and fire ants of different sizes showed
a trend for more crazy ants than fire ants to be involved in fights (Figure 2.2) and higher
crazy ant than fire ant mortality after one hour (Figure 2.3). Both ant species experienced
higher mortality after exposure to the other species than either species experienced in
21
isolation. Fire ant worker size had little effect on antagonistic interactions with crazy
ants, as aggression score, the number of crazy ants fighting, and mortality for both
species were not influenced by fire ant size. The exception is that small fire ants fought
more than 50% more often than large fire ants (Figure 2.1). The tendency of small fire
ant workers to fight more often may be biologically important, as small fire ants are
nearly 3 times more common than large fire ants based on the average distribution of
worker sizes in our field-collected colonies (Table 2.1). This may mean that crazy ants
are more likely to encounter small workers that readily engage in fights than large
workers that avoid fights, especially since smaller fire ant workers have been found ot
forage farther from nests (Martin and Vinson 2008). The increase in fire ant fighting
does not correlate with an increase in mortality for either crazy ants or fire ants,
suggesting that the likelihood of engaging in fights may not be indicative of the outcome
of interactions. Overall, crazy ants, which spray formic acid with an acidopore, had an
average mortality nearly twice that of fire ants (Figure 2.3), which inject acid using a
stinger (Holldobler and Wilson 1990). This large average difference was not statistically
significant perhaps due to the small sample size and the low power of a non-parametric
test, and it is possible that the difference in mortality between the two species is
biologically significant. Though spraying acid can be a very effective method of
attacking other insects, as the acid is absorbed through the exoskeleton, crazy ants may
have less success in fights with fire ants due to the difference in size, as even small fire
ant workers are significantly larger than crazy ants. If the advantage of fire ants over
crazy ants in individual clashes is biologically significant, it could be important in
22
limiting the establishment and spread of crazy ants in habitats where fire ants are
common.
Diet differences had no effect on intraspecific crazy ant interactions after 12 days.
No fighting was observed between any of the workers, and the highest scoring interaction
recorded was avoidance, which occurred only 5 times. Mortality did not differ among
pairs that received different diets, pairs receiving the same diet, and controls of 5 ants.
The lack of antagonistic interactions between crazy ants that received different diets may
indicate that diet changes alone do not disrupt the chemical profile of the workers enough
to overcome the unicoloniality of the introduced population. However, several studies
that have examined the effect of diet on intraspecific interactions have found that
aggression can be induced between unicolonial nests or former nestmates (Liang et al.
2001, Silverman and Liang 2001, Buczkowski and Silverman 2006, Corin et al. 2007).
For example, Corin et al. (2007) found aggression between unicolonial Argentine ant
nests after 56 days of diet maintenance, and Silverman and Liang (2001), who studied
interactions between separated nestmates consuming different prey at several intervals for
56 days, found that former nestmates fought each other after 28 days. Additionally, Lim
et al. examined interactions between nestmates of Paratrechina longicornis and found
antagonistic behavior beginning at 21 days after the implementation of diet treatments
(Lim et al. 2003). Together, these studies could suggest that the absence of intraspecific
aggression in Rasberry crazy ants is due to the treatments not being maintained long
enough to for nestmate recognition to be affected. Nevertheless, one study of Argentine
ants showed antagonistic behavior between nestmates after as little as 2 minutes of
contact with a prey item, the brown-banded cockroach, Supella longipalpa (Liang et al.
23
2001). Therefore, changes in nestmate recognition are therefore conceivable within 12
days, and the protein sources used in the diets, though from different orders of insects,
may not have had distinct enough chemical profiles to affect nestmate recognition.
Alternatively, crazy ant nestmate recognition may be based primarily on genetic cues
which may not be significantly affected by environmental factors like diet, as is the case
with Argentine ants (Suarez et al. 2008 and references therein). Ultimately, further study
using longer periods of diet maintenance and different protein types, such as cockroaches,
may affect nestmate recognition in Rasberry crazy ants, though this study did not reveal
any evidence of inducible intraspecific aggression. As the density of colonies increases,
this lack of intraspecific aggression may allow continued population growth.
In keeping with the results of the intraspecific aggression assays, diet had only a
limited effect on interactions between crazy ants and fire ants. Crazy ants receiving a low
sugar diet engaged in higher scoring aggressive interactions with fire ants than crazy ants
on a high sugar diet (Figure 2.4). Yet sugar level did not affect the number of ants
fighting or mortality for either species. Protein type had no effect on antagonistic
interactions, which is consistent with the findings of Alloway et al. (1991), who found no
difference in the number of fighting ants of two species of Leptothoracine ants that were
fed diets that differed in protein type and vitamin and mineral source but not sugar
source. Overall, the two species did not differ significantly in the average number of
workers fighting or in mortality after one hour, and there was no effect of diet on the
difference between the two species.
Carbohydrates play an important role in ant activity levels (Grover et al. 2007).
Therefore, it is expected that sugar level would affect interspecific antagonistic
24
interactions. What may seem surprising is that the data show that crazy ants receiving
fewer carbohydrates engage in more aggressive interactions with fire ants (Figures 2.3).
However, in a study examining the effect of diet on intraspecific interactions in Argentine
ants, Grover et al. (2007) found that aggression was decreased in the absence of sugar,
but there was no difference in aggression scores between low sugar and high sugar
treatments. Furthermore, per capita activity (based on the exploration of a structure not
containing food) was higher in the low sucrose treatment than in either the sucrose-free
treatment or the high sucrose treatment. In light of these findings, it is conceivable that
crazy ants that are sugar-limited may experience increased per capita activity and
therefore engage in more intense fights with another species, in this case fire ants. The
fact that there was no effect of sugar level on mortality suggests that fighting intensity is
not closely correlated with mortality for Rasberry crazy ants. Overall, the finding of
increased aggression score on a low sugar diet provides new insight in the behavior of
crazy ants, showing a similarity to the highly invasive Argentine ant (Grover et al. 2007).
The findings of this study give insights on the biotic factors affecting the spread
of the introduced Rasberry crazy ant populations. The most common ant in the
introduced range, the red imported fire ant, has less mortality when facing off in fights
with crazy ants. Additionally, small fire ants, which are most common, engage in fights
more often than larger fire ants. Together, these results suggest that fire ants may have a
competitive advantage over crazy ants and may be able to defend their territory from a
neighboring colony of crazy ants, though colony-level studies will put the individual-
level advantage in the context of relative abundance and worker activity. On the other
hand, when crazy ants consume a reduced sugar diet, they fight more intensely, without a
25
corresponding increase in mortality. Therefore, a crazy ant colony that has been
competitively excluded from sugar sources by fire ants or other ants may engage intense
fights, inflicting various degrees of mortality on other ant species without suffering high
mortality themselves. Combined with the fact that differences in diet showed no sign of
affecting nestmate recognition, large supercolonies of sugar-limited crazy ants may
challenge fire ant dominance in certain areas by overwhelming fire ant nests with
considerable numbers of very active workers that readily engage in fights. Nevertheless,
because fire ants have almost half as much mortality as crazy ants in individual
interactions, fire ant colonies could provide biotic resistance to the crazy ant invasion in
places where fire ants outnumber crazy ants and can afford to engage in fights without
serious losses in workers or foraging efficiency. In sum, these findings help explain the
occurrence of large supercolonies of crazy ants but suggest that established fire ants may
provide some biotic resistance to crazy ants, especially if crazy ants are not sugar-limited.
Further studies on colony-level interactions will further elucidate the ability of an
established invasive species to resist the spread of a newly introduced species.
Acknowledgements
I would like to thank Matt Barajas, Juli Carrillo, Marilu Corona, Christopher
Gabler, Summer Jones, Max Quintanilla, Evan Siemann, and Erica Soltero for their field
and laboratory assistance. Financial support was provided by a Wray-Todd Fellowship to
KH.
26
Literature Cited
Alloway, T. M, A. Leighl, and D. Ryckman. 1991. Diet does not affect intercolonial
fighting in Leptothoracine ants. Insectes Sociaux 38:189-193.
Andersen, A. N. and A. D. Patel. 1994. Meat ants as dominant members of Australian ant
communities - an experimental test of their influence on the foraging success and
forager abundance of other species. Oecologia 98:15-24.
Buczkowski, G., R. Kumar, S. L. Suib, and J. Silverman. 2005. Diet-related modification
of cuticular hydrocarbon profiles of the Argentine ant, Linepithema humile,
diminishes intercolony aggression. Journal of Chemical Ecology 31:829-843.
Buczkowski, G. and J. Silverman. 2006. Geographical variation in Argentine ant
aggression behaviour mediated by environmentally derived nestmate recognition
cues. Animal Behaviour 71:327-335.
Buren, W. F., G. E. Allen, W. H. Whitcomb, F. E. Lennartz, and R. N. Williams. 1974.
Zoogeography of imported fire ants. Journal of the New York Entomological
Society 82:113-124.
Corin, S., K. Abbott, P. Ritchie, and P. Lester. 2007. Large scale unicoloniality: the
population and colony structure of the invasive Argentine ant {Linepithema
humile) in New Zealand. Insectes Sociaux 54:275-282.
Davidson, D. W. 1997. The role of resource imbalances in the evolutionary ecology of
tropical arboreal ants. Biological Journal of the Linnean Society 61:153-181.
Finke, M. D. 2002. Complete nutrient composition of commercially raised invertebrates
used as food for insectivores. Zoo Biology 21:269-285.
27
Grover, C, A. Kay, J. Monson, T. Marsh, and D. Holway. 2007. Linking nutrition and
behavioural dominance: carbohydrate scarcity limits aggression and activity in
Argentine ants. Proceedings of the Royal Society B 274:2951.
Holway, D., L. Lach, A. Suarez, N. Tsutsui, and T. Case. 2002. The causes and
consequences of ant invasions. Annual Review of Ecology, Evolution, &
Systematics 33:181-233.
Holway, D., A. Suarez, and T. Case. 1998. Loss of intraspecific aggression in the success
of a widespread invasive social insect. Science 282:949.
Howard, R. W. and G. J. Blomquist. 2005. Ecological, behavioral, and biochemical
aspects of insect hydrocarbons. Annual Review of Entomology 50:371-393.
Holldobler, B. and E. O. Wilson. 1990. The ants. Springer-Verlag, Berlin, Germany.
Kaspari, M. and M. Weiser. 1999. The size-grain hypothesis and interspecific scaling in
ants. Functional Ecology 13:530-538.
Liang, D., G. Blomquist, and J. Silverman. 2001. Hydrocarbon-released nestmate
aggression in the Argentine ant, Linepithema humile, following encounters with
insect prey. Comparative Biochemistry and Physiology, Part B 129:871-882.
Lim, S. P., A. S. C. Chong, and C. Y. Lee. 2003. Nestmate recognition and intercolonial
aggression in the crazy ant, Paratrechina longicornis (Hymenoptera: Formicidae).
Sociobiology 41:295-305.
Lowe, S., M. Browne, S. Boudjelas, and M. De Poorter. 2004. 100 of the world's worst
invasive species: a selection from the Global Invasive Species Database. Invasive
Species Specialist Group, Gland, Switzerland.
28
Mack, R. N., D. Simberloff, W. M. Lonsdale, H. Evans, M. Clout, and F. A. Bazzaz.
2000. Biotic invasions: Causes, epidemiology, global consequences, and control.
Ecological Applications 10:689-710.
Martin, J. B. and S. B. Vinson. 2008. The influence of travel distance on sugar loading
decisions and water balance in the central place foraging ant Solenopsis invicta.
Insectes Sociaux 55:129-136.
Menke, S., R. Fisher, W. Jetz, and D. Holway. 2007. Biotic and abiotic controls of
Argentine ant invasion success at local and landscape scales. Ecology 88:3164-
3173.
Meyers, J. M. 2008. Identification, distribution and control of an invasive pest ant,
Paratrechina sp. (Hymenoptera, Formicidae), in Texas. PhD Thesis. Texas A&M
University, College Station, TX.
Mitchell, C. E. and A. G. Power. 2003. Release of invasive plants from fungal and viral
pathogens. Nature 421:625-627.
Ness, J. and J. Bronstein. 2004. The effects of invasive ants on prospective ant mutualists.
Biological Invasions 6:445-461.
Pimentel, D., L. Lach, R. Zuniga, and D. Morrison. 2000. Environmental and economic
costs of nonindigenous species in the United States. Bioscience 50:53-65.
Ricklefs, R. E. 2001. The economy of nature. 5th edition. W. H. Freeman and Company,
New York, NY.
Roulston, T., G. Buczkowski, and J. Silverman. 2003. Nestmate discrimination in ants:
effect of bioassay on aggressive behavior. Insectes Sociaux 50:151-159.
29
Savolainen, R. and K. Vepsalainen. 1988. A competition hierarchy among boreal ants -
impact on resource partitioning and community structure. Oikos 51:135-155.
Silverman, J. and D. Liang. 2001. Colony disassociation following diet partitioning in a
unicolonial ant. Naturwissenschaften 88:73-77.
Sorvari, J., P. Theodora, S. Turillazzi, H. Hakkarainen, and L. Sundstrom. 2008. Food
resources, chemical signaling, and nest mate recognition in the ant Formica
aquilonia. Behavioral Ecology 19:441-447.
Suarez, A. N., D. A. Holway, and N. D. Tsutsui. 2008. Genetics and behavior of a
colonizing species: the invasive Argentine ant. American Naturalist 172:S72-S84.
Suarez, A., N. Tsutsui, D. Holway, and T. Case. 1999. Behavioral and genetic
differentiation between native and introduced populations of the Argentine ant.
Biological Invasions 1:43-53.
Torchin, M. E. and C. E. Mitchell. 2004. Parasites, pathogens, and invasions by plants
and animals. Frontiers in Ecology and the Environment 2:183-190.
Walters, A. and D. Mackay. 2005. Importance of large colony size for successful
invasion by Argentine ants (Hymenoptera: Formicidae): Evidence for biotic
resistance by native ants. Austral Ecology 30:395-406.
Wang, D., Y. Y. Bai, J.H. Li, and C. X. Zhang. 2004. Nutritional value of the field
cricket. Insect Science 11:275-283.
Williams, D. F. 1994. Control of the introduced pest Solenopsis invicta in the United
States.w D. F. Williams, editor. Exotic ants: biology, impact, and control of
introduced species. Westview Press, Inc., Boulder, Colorado.
30
Table 2.1. Head lengths of fire ants and crazy ants and fire ant colony size distribution.
Ants that do not share the same level letter significantly differ in head length (as
indicated in a Tukey's post hoc test). Average # is the number of fire ants in each size
class in a haphazardly selected sample of 21 to 36 ants from 10 colonies. Percentages are
from an average sample size of 28.36 ants. Errors are ± 1 SE.
Ant Length (mm) Level Average # % of Sample Large fire 1.28 ±0.02 A 4.73 ±5.02 16.1% Medium fire N/A N/A 10.91 ±7.46 37.2% Small fire 0.78 ±0.02 B 12.72 ±5.48 46.7% Crazy 0.69 ± 0.01 C N/A N/A
31
Figure 2.1. The effect of fire ant worker size on the average number of fire ants fighting
during each observation (out of a total of 5 fire ants). Small fire ants fought more often
than large fire ants (p = 0.040). Error bars indicate + 1 SE.
O) _c '•*-> ^z O)
M—
(/) •*->
CO
9 !_ M =
o l _ 0) J2 E S2
1.6 -I
1.4 -
1.2 -
1.0 -
0.8 -
0.6 -
0.4 -
0.2 -
o.o J 1 . ' ' . —
Large Small
Fire ant size
Figure 2.2. The number of Rasberry crazy ants and fire ants engaged in fights. On
average, crazy ants had more workers engaged in fights than fire ants did in aggression
trials between the two species (p = 0.063). Error bars indicate + 1 SE.
1.4 •
D) 1 2 -C
fight
i o
1
•*-• 0.8 -C (0
* 0 0.6 -
(D - | 0.4-
Z 0.2 -
T
Crazy ants Fire ants
Species
33
Figure 2.3. The number of dead Rasberry crazy ants and dead fire ants after 1 hour.
Crazy ants had higher mortality than fire ants (p = 0.039). The maximum possible
mortality was 5 ants. Error bars indicate + 1 SE.
3.0
C/> 2 5
c T3 2.0 03 <D
"O M_ 1.5
o 1— <D
-Q 1.0
E ^ 0.5 -
0.0
Crazy ants Fire ants
Species
Figure 2.4. The effect of sugar level in crazy ant diet on the average aggression score for
each set of fire ants and crazy ants. Aggression scores are assigned according to the scale
described by Suarez et al. (1999). Aggression scores for interactions between fire ants
and crazy ants were higher when crazy ants received a low sugar diet (p = 0.037). Error
bars indicate + 1 SE.
1.0 -
1.6 -
<D 1 4 "
O 1.2 • CO C 1.0 -
o
"8 08
O ) 0.6-
< 0.4-
0.2 -
T
T
High Low
Sugar level in crazy ant diet
35
Chapter 3
Colony-Level Competition Between Rasberry Crazy Ants {Paratrechina sp. nr. pubens) and Red Imported Fire Ants {Solenopsis invicta)
Abstract
In further examine whether red imported fire ants (Solenopsis invicta) can provide
biotic resistance to Rasberry crazy ants (Paratrechina sp. nr. pubens), I conducted a
laboratory experiment and a field experiment that examined colony-level competition
between these two species. In the laboratory experiment, colonies of fire ants and crazy
ants were standardized by worker number or by biomass and used as either controls or in
competing pairs. Control colonies had access to a foraging tray with sugar water and
mealworms, while competing colonies shared a foraging tray with a colony of the
opposing species. For both colony sizes, and in both control and competition settings,
crazy ants recruited more quickly and in higher abundances to baits. Crazy ants also
controlled significantly more baits than fire ants did at the end of the experiment. These
results suggest that crazy ants, when evenly matched in colony size with fire ants, are
capable of discovering baits first and excluding fire ants from resources, thus breaking
the dominance-discovery trade-off. In order to simulate the introduction of crazy ants to
new areas, I also conducted a field experiment that moved small crazy ant colonies from
a heavily invaded site to another location 750m away that had not yet been invaded. I
then monitored the ant recruitment to sugar and protein baits for three weeks. Crazy ants
discovered baits more quickly, but were subsequently displaced by fire ants, which
recruited to both sugar and protein in higher abundances than crazy ants and controlled
the majority of baits. Crazy ant recruitment increased through time in crazy ant addition
36
locations, but fire ant recruitment decreased in both control and crazy ant addition
locations, suggesting that crazy ants responded to, rather than caused, drops in fire ant
recruitment. Together, these findings suggest that crazy ants may not be able to displace
fire ants in areas that cannot support high crazy ant densities, a factor which may restrict
the potential spread of crazy ant populations and enable the two invasive species to
coexist in a patchy distribution.
Introduction
Colony-level antagonistic interactions are fundamental to the structuring of ant
communities (Holldobler and Wilson 1990). Though aggression assays which focus on
worker interactions usually accurately predict whether two colonies will behave
aggressively towards one another (Roulston et al. 2003), they do not necessarily predict
which colony will have a competitive advantage. In some species, such as the highly
invasive Argentine ant (Linepithema humile), an individual ant may not be successful in
one-on-one interactions with their competitors, but as part of a larger colony, the ants can
outcompete and displace competing colonies (Holway 1999, Buczkowski and Bennett
2008). In this study, a laboratory and a field experiment were used to assess the colony-
level competitive ability of the non-native Rasberry crazy ant, Paratrechina sp. nr.
pubens. In the lab, Rasberry crazy ants were paired with red imported fire ants,
Solenopsis invicta, while crazy ant colonies in the field interacted with the existing ant
community in the coastal tallgrass prairie of the Texas Gulf Coast, a habitat which is
dominated by fire ants. This is the first time that colony-level competitive ability has
been assessed in the recently introduced invasive threat, Rasberry crazy ants, and it is the
37
first study, to my knowledge, examining colony-level competition between a long
established invasive ant and newly introduced ant species.
Relative colony size has been shown to be important in determining the outcome
of competitive interactions between colonies. In Argentine ants, competitive
performance increases with colony size in laboratory trials against native ants, and a
threshold of relative abundance is necessary for Argentine ants to maintain a dominant
worker presence at baits (Holway and Case 2001, Walters and Mackay 2005). Studies
with red imported fire ants have also indicated the importance of relative abundance, as
competitive success is affected by colony size in trials with native as well as invasive ants
(Morrison 2000, Kabasbima et al. 2007). Therefore, in the laboratory experiment we
paired crazy ants and fire ants in colony units that had been standardized by either
biomass or worker number. Additionally, the crazy ant colonies used in the field
experiment were smaller than the existing fire ant nests. Together, testing the
competitive ability of these different relative abundances of crazy ants may lead to
important conclusions about the ability of the Rasberry crazy ant to expand its range in
the face of competition from fire ants.
One advantage of a colony-level approach is that is possible to examine
recruitment to and dominance of baits, in addition to fighting and mortality. Ant species
are often classified as either opportunists, which specialize in discovering baits quickly;
extirpators, which arrive to baits later but dominate discovered resources; or insinuators,
which inconspicuously thieve food from dominant ants (Wilson 1971). Fellers (1987)
first proposed that there may be an evolutionary trade-off between resource discovery and
behavioral dominance of resources, a hypothesis known as the dominance-discovery
38
tradeoff. However, more recent studies have shown that some invasive ant species can
break this trade-off in their introduced range and can both discover and dominate baits
(Davidson 1998). Therefore, examining the colony-level competitive interactions of
crazy ants and fire ants both in the laboratory and in the field may lead to new insight into
the community role of crazy ants in their introduced range, and more specifically,
whether crazy ants break the dominance-discovery trade-off.
Our experimental design allowed us to ask the following questions: 1) Are fire
ants or crazy ants more successful in colony-level interactions? 2) Does relative
abundance of crazy ants affect the outcome of colony-level interactions? 3) Do crazy ants
follow the dominance-discovery tradeoff? and 4) Do crazy ants fill the same role in the
dominance-discovery tradeoff in field conditions when they are much less abundant than
fire ants? The answers to these questions may lead to a greater understanding of the
Rasberry crazy ant introduction, as well as the ability to better predict the future of
Rasberry crazy ant populations in their introduced range. More importantly, these
questions examine whether fire ants, an well-established invasive species, can provide
biotic resistance to the spread of the newly introduced crazy ants.
Methods
Colony-level competition
All ants used in this experiment were from field-collected, polygynous colonies.
Paratrechina sp. nr. pubens colonies were collected in Pearland, TX (29.55°N, 95.28°W)
on May 24 and May 31,2008. Solenopsis invicta colonies were collected in Katy, TX
(29.93°N, 95.94°W) in early May 2008. After separating ants from the nesting material
39
with which they were collected by flooding the nests, each colony was placed in a 24cm
x 1 lcm plastic nest box. To prevent ants from escaping, nest boxes were ringed with a
thin layer of Tanglefoot (Tanglefoot, Grand Rapids, Michigan, USA) near the top of the
inside walls. Each colony was fed a standard diet of sugar water daily and half of a
freezer-killed mealworm (Armstrong's Cricket Farm, West Monroe, LA) every other day.
In order to obtain colonies of uniform size for competitive pairings, field-collected
colonies were divided into smaller units from June 18 to July 1,2008. Each unit
contained one queen, no brood, and either 1000 workers in colonies matched by worker
number or 500 mg of workers in colonies matched by biomass. Because counting
individual workers was not feasible, we used an average per worker mass to extrapolate
the mass of 1000 workers. For crazy ants, we randomly selected and weighed between
16 and 26 workers from each of 10 different colonies, recording the number of workers
and the total weight each time. For fire ants, we repeated this process with 22 to 28
workers from each of 13 different colonies. The average weight per Rasberry crazy ant
was 0.68mg, and the average weight per fire ant was 1.43mg. Therefore for colonies that
were matched by worker number, we used 680.0 ± 50.0mg of crazy ants or 1430.0 ±
lOO.Omg of fire ants.
In total, we created 16 units of crazy ants with 500mg of workers, 16 units of
crazy ants with 1000 workers, 15 units of fire ants with 500mg of workers, and 16 units
of fire ants with 1000 workers. These colony units were haphazardly assigned to a
treatment of either control or competition such that there were 8 control colony units of
crazy ants standardized by biomass, 8 control colony units of crazy ants standardized by
worker number, 7 control colony units of fire ants standardized by biomass, 8 control
40
colony units of fire ants standardized by worker number, 8 competition pairs of crazy ant
and fire ant colony units standardized by biomass, and 8 competition pairs of crazy ants
and fire ants standardized by worker number.
For all but two units, the queen and the workers came from the same source
colony. The two exceptions were crazy ant units from a source colony with a shortage of
queens in relation to the amount of workers. Queens from a different source colony were
used, however all of crazy ant source colonies came from a population where the ants
behave as a supercolony (H6lldobler and Wilson 1990), and there was no aggression
toward the adopted queen. Most (14 of 20) fire ant source colonies were used only once
to make a colony unit, though 6 of 20 were used between 2 and 5 times. Each field-
collected crazy ant colony was used to make between 2 and 6 colony units. In 5 cases,
the same source colony was used to make two colony units for the control treatment with
the same standardization factor (500 mg workers or 1000 workers), and in one case the
same source colonies were used for 2 competing pairs of fire ants and crazy ants. We
treated these colony units and pairs as independent data points because the queen and
workers were unique for each unit and a preliminary analysis showed that the response
variables of these units were not more similar to each other than to other colony units.
Colony units were housed in 20cm x 12cm x 6cm plastic containers with a layer
of Tanglefoot around the upper edge. Each colony unit was allowed to acclimate to its
container for 2 to 14 days before the start of the experiment, during which time all colony
units continued the diet of sugar water and freezer-killed mealworms. Colony units were
then starved for 24 hours before the start of the experiment on July 3,2008, at which
point they were given access to an identical foraging tray via a 5cm-wide bridge of
41
uncoated metal hardware cloth. Each foraging tray contained a freezer-killed mealworm
(cut in half) and a 1-ml Eppendorf tube filled with sugar water-soaked cotton placed on
top of a 7.6cm x 12.7cm white card. Each colony unit in the control treatment was
connected to a unique foraging tray, while each pair of competing colony units shared a
foraging tray. The set of hardware cloth bridges attaching competing colonies to the
foraging tray also allowed for the movement of ants between nest boxes. Each bait was
removed, weighed, and replaced with a fresh bait item on days 3 and 5 of the experiment.
The average desiccation of control baits (without ant access) was subtracted from the
difference between the initial and final weight of each bait item to determine the mass of
bait consumed during the 48-hour period. This calculation method sometimes led to
negative values for overall resource consumption.
Behavioral data were collected every 15 minutes for 1 hour and 45 minutes
beginning 10 minutes after the bridges were first put in place at 1400h CST on July 3,
2008. Data was then collected each hour until 3 hours and 55 minutes after the start of
the experiment and at 24-hour intervals from the start of the experiment until its
completion on day 6 (July 8,2008). On both days 3 and 5, behavioral data were also
collected after sugar water and mealworms were replaced, at 51 hours and 10 minutes
and 99 hours and 10 minutes after the start of the experiment. During each data
collection, the number of ants of each species at the sugar water and at the mealworm
was recorded for all colony units. In addition, for paired competition units, we recorded
the number of ants of each species involved in fights, as well as the number of ants in the
nest box of the competing species. After the last data collection event at 1400h on July 8,
42
2008, bridges were removed and dead ants were collected from nest boxes and foraging
trays and stored in alcohol until they were counted to assess mortality from each colony.
Statistical analysis
All statistical analyses for the colony-level competition experiment were
performed in JMP 7.0.2 (SAS Institute Inc., Cary, NC, 2007). The following response
variables were used in analyses for each colony or pair of colonies (in the case of
resource consumption): mortality at the end of the experiment (number of dead ants not
stuck in Tanglefoot), mealworm and sugar water consumption (averaged across the 3
pieces of each bait given to each colony), total recruitment across all time periods and
time until recruitment (when 2 or more ants of the same species were present) for each
bait type, the total number of ants fighting and the total number of ants in the other
species' nest box summed across all observations, and final bait control (the species
present on each bait type after food was replaced on day 5). We tested for correlations
between all response variables. For all variables other than bait dominance, the data were
analyzed as follows: control colonies were compared using ANOVA's with species,
standardizing factor (biomass or worker numbers), and the interaction between the two as
factors. Post-hoc analyses on significant interactions were conducted using Tukey's HSD
tests. To compare between species in the competition treatment, the difference between
the two species was calculated and the mean was tested for significant deviation from 0
using a 1-group, 2-tailed t-test. The effect of standardizing factor was analyzed using a
one-way ANOVA. The following transformations were performed to meet assumptions
of normality: mealworm consumption for control colonies and total ants fighting were
square root transformed, and mortality and mealworm and sugar water recruitment were
43
logio transformed. Three outlying data points were excluded from analysis: one crazy ant
control colony for sugar water consumption, one crazy ant control colony for mortality,
and 1 competing pair of colonies for total ants in the other species nest box. When
transformations to normality were not possible, Wilcoxon signed rank tests were
performed for each factor (instead of an ANOVA) or mean difference (instead of a 2-
tailed t-test). The specific test used for each response variable is shown in Table 3.1.
Bait dominance was analyzed for a species effect using binomial probability.
Crazy ant addition field experiment
Field research was conducted at the University of Houston Coastal Center
(UHCC) in La Marque, TX (29.38°N, 95.04°W). The 400-hectare property has
approximately 80 hectares of undisturbed native prairie, with the remainder of the
property composed primarily of non-native woody species (especially Chinese tallow tree
[Triadica sebifera]). In April 2007, Rasberry crazy ants were found in high abundances
in the southwest corner of the UHCC property. It is unknown how the ants were
introduced to UHCC; they may have arrived with a load of contaminated fill dirt, or they
may have spread from a possible infestation at the landfill on the western border of the
UHCC property. In the infested area, nests were located in the woods, but foraging ants
were readily apparent in open spaces, including roadways and prairie. The population
appeared to be unicolonial, with no noticeable aggression between nests (Holldobler and
Wilson 1990).
Because the high-density infestation was limited to the southwest corner of the
property, we were able to locate an area without a high abundance of crazy ants in the
southeast corner of UHCC in a prairie bordered by woods, approximately 750m from the
44
infestation. We selected 24 locations at the edge of the prairie and the woods, each 25m
apart. Each location was randomly assigned to either a crazy ant colony addition
treatment or control. Prior to establishing treatments, we baited for ants on June 17,
2009, following the same protocol used in all subsequent baitings. At each location, we
placed two 9cm diameter Petri dishes; one dish contained a ~2gm slice of Libby's Vienna
Sausage, and the other dish held a cotton ball soaked with 30% honey water. We
monitored each Petri dish for 10 minutes or until it was discovered, at which point we
recorded the ant species and the number of minutes that had passed since opening the
bait. One hour after baits were opened, we closed the Petri dishes and placed them in a
freezer until the number of each species in each dish could be recorded.
Five ant species were collected during the initial baiting: Solenopsis invicta (17
dishes), P. sp. nr. pubens (8 dishes), Solenopsis sp. cf. geminata (3 dishes),
Crematogaster sp. (1 dish), and Pheidole sp. (1 dish). Solenopsis invicta was the most
commonly collected ant species and had the highest average recruitment with 153.2 ants
per dish. Crazy ants recruited to a total of 7 locations: to sugar baits at 5 locations, to
protein at 1 location, and to both sugar and protein at 1 location. The average crazy ant
recruitment was 25 ants, and all but the 2 protein baits had 12 or fewer crazy ants.
On June 19,2009,12 crazy ant nests were collected in decaying logs from the
high-density infestation area. Collection of superficially similar nests from a different
site in 2008 revealed that these nests often contained more than 3000 workers and at least
5 queens (personal observation, K. Horn). Therefore, while we were unable to measure
the size of the experimental colonies, each colony likely had at least 1000 workers and
more than 2 queens. Because the nests had been collected from a wooded area, we
45
placed a nest of crazy ants 1.5m away from the edge of the prairie into the woods at each
location assigned to the crazy ant addition treatment. At each control location, we placed
a similar amount of ant-free decaying wood 1.5m into the woods. After the treatments
were established, ant baiting was conducted on June 24, June 29, July 3, July 8, July 14,
and July 29 which was 5, 10,14,19,25, and 40 days after colony additions.
Statistical analyses
Crazy ants and fire ants were the only two species that recruited to baits with
regularity and abundance, therefore the analyses focused on these two species. On all
days except July 29, baiting data revealed higher average crazy ant recruitment at crazy
ant addition locations than at control locations. Therefore, because the treatment was not
reflected in census data on July 29, that date was excluded from analysis. Repeated
measures ANOVA's performed in StatView 5.0 (SAS Institute Inc., Cary, NC, 1998)
were used to determine if crazy ant and fire ant recruitment to baits after 1 hour changed
through time at control and crazy ant addition locations using the sampling dates from
June 24 to July 14 (i.e. after crazy ants were added). These ANOVA's included
treatment and bait type, as well as all possible interactions, as factors. Using presence
and absence at baits after 1 hour on July 8 and July 14, we used 2 contingency tables for
each bait type to determine if a single ant species dominated baits more often than 2 or
more ant species shared baits. To examine the number of baits discovered within 10
minutes, we used a series of 2x2 contingency tables tested for significance in Microsoft
Excel or in JMP 7.0.2. Specifically, we tested whether the number of baits discovered in
10 minutes differed by bait type, species, or treatment. We also tested whether average
discovery time differed by species, bait type, treatment, or any of the interactions of those
46
variables using an ANOVA in SAS 9.1 (SAS Institute Inc., Cary, NC, 2003). Because
only 5-10 of 48 baits were discovered within 10 minutes on any given day, all dates were
summed for contingency tables and recruitment time analysis. None of the locations had
repeated discoveries of sugar baits, but 10 out of 19 locations that discovered protein
dishes had repeated discoveries. Though repeated discoveries of baits on different days
are not truly independent data points, this was the only way to conduct meaningful
statistical tests.
Results
Colony-level competition
For control colonies, ant species was often a significant factor. Crazy ants had
significantly lower mortality than fire ants (Table 3.1, Figure 3.1) and consumed
significantly less sugar water (Table 3.1, Figure 3.2), but there was no difference between
species in mealworm consumption (Table 3.1). Crazy ants recruited in greater numbers
than fire ants to mealworms but not to sugar water (Table 3.1, Figure 3.3). Crazy ants
also discovered and recruited to both mealworms and sugar water more quickly than fire
ants did (Table 3.1, Figure 3.4).
Standardizing factor (mass versus number of workers) generally did not
significantly affect response variables for control colonies (Table 3.1). The exception
was that mealworm and sugar water total recruitment were both higher for colony units
standardized by worker number (worker numbers were higher for colonies standardized
by worker number than by biomass). The interaction between species and standardizing
factor was significant only for total recruitment to sugar water (Table 3.1). For sugar
47
water, recruitment was significantly higher for both species in the worker number
treatment than for fire ants in the biomass treatment, but not significantly higher than
crazy ants in the biomass treatment. The sugar water recruitment in the biomass
treatment was not significantly different for the two species.
For competition colonies, species was again frequently a significant predictor.
Differences in recruitment abundances and recruitment and discovery times were also
strongly affected by species identity, as crazy ants had a higher total recruitment to both
sugar water and mealworms and recruited to sugar water and mealworms more quickly
than did fire ants (Table 3.1, Figures 3.3 and 3.4). Crazy ants also had higher numbers of
ants engaged in fights and in the other species' nest box than did fire ants (Table 3.1,
Figure 3.5). However, mortality was not different between the two species in
competition (Table 3.1, Figure 3.1). Crazy ants were the only species at baits at the end
of the experiment significantly more often than fire ants (Table 3.1). Standardizing factor
was not significant for any of the response variables for competition colonies (Table 3.1).
All significant correlations between response variables are shown in Table 3.2. In
general, discovery and recruitment times between and within bait types were positively
correlated, and sugar water and mealworm recruitment abundances were positively
correlated. Sugar water and mealworm recruitment times were negatively correlated with
the recruitment abundances for these baits. Mealworm recruitment time was also
negatively correlated with mortality.
Crazy ant addition field experiment
On average, crazy ant recruitment increased through time, but not significantly so
(F4,i76= 1.683, p = 0.1559, Figure 3.6). There was a trend for crazy ant recruitment to be
48
higher in crazy ant addition locations than control locations (Fi,44= 3.697, p = 0.0610),
and crazy ant recruitment was higher for protein than for sugar (Fi(44 = 6.770, p = 0.0126,
Figure 3.7). The interaction between bait type and treatment was sizeable but not
significant (Fi;44 = 2.825, p = 0.0990). There were no significant interactions with time
and any other factors. Fire ant recruitment decreased significantly through time (F4J76 =
5.154, p = 0.0006, Figure 3.6), but recruitment was not affected by treatment (Fi,44 =
0.026, p = 0.8718). Recruitment to protein baits was much higher than recruitment to
sugar baits (Fi^ = 22.817, p < 0.0001, Figure 3.7). On average, per bait recruitment was
higher for fire ants than for crazy ants (Figure 3.6). Fire ants also controlled significantly
more protein baits (p = 0.0394) but not sugar baits (p = 0.1444) after 1 hour (on July 8
and 14). On both July 8 and July 14, there were no instances of crazy ants and fire ants
co-occurring at a protein bait (x = 10.1931, p = 0.0006), yet there was no evidence of
sugar bait exclusion by either species (x = 0.9470, p = 0.9538).
Summing all time periods, crazy ants discovered 16 of the baits in 10 minutes,
and fire ants discovered 15 of the baits. Protein baits were discovered in 10 minutes
significantly more often than sugar baits, regardless of species (x = 11.242, p = 0.0008).
Fire ants and crazy ants did not differ in the amount of protein baits (p = 0.1489) or sugar
baits (p = 0.2734) they discovered. Protein baits were discovered more often than sugar
baits by both crazy ants (x = 7.185, p = 0.0074) and fire ants (x = 3.615, p = 0.0573).
Treatments did not affect the number of baits discovered by fire ants (x = 1.810, p =
0.1785) or by crazy ants (x = 2.467, p = 0.1163). Similarly, treatment had no effect on
average discovery time (Fi,34 = 0.26, p = 0.6148) for the two species, and the treatment by
species interaction was insignificant (Fi>34= 0.03, p = 0.8742). In contrast, bait type
49
significantly affected discovery time (Fi^ = 6.71, p = 0.0140), with sugar baits
discovered more quickly than protein baits (Figure 3.8). Species was also a significant
factor (F134 = 7.62, p = 0.0092), as crazy ants discovered baits more quickly than fire ants
(Figure 3.8).
Discussion
The results of the colony-level competition experiment indicate that crazy ants
recruit to baits more quickly than fire ants both in isolation and in competition (Figure
3.4). However, despite their quick recruitment, they do not consume more resources than
fire ants in control colonies (Table 3.1). In fact, the only significant difference between
species in resource consumption was for sugar water consumption, of which fire ants
consumed more (Figure 3.2). Yet, crazy ants were the species found most often at both
bait types at the end of the experiment in competition colonies (Table 3.1), suggesting
that their recruitment times and abundances may be important in maintaining access to
baits when faced with a competitor.
In general, significant differences between species in controls remained
significant for the same response variables in competition colonies (Table 3.1). Crazy
ants recruited more quickly to mealworms and sugar water and in higher numbers to
mealworms than fire ants did in both control and competition colonies (Figures 3.3 and
3.4). In fact, the crazy ant advantage was larger in competition than in control. This
finding suggests that crazy ants, which have an inherent advantage over fire ants in
recruitment speed and numbers, may have deterred fire ants from discovering and
recruiting to baits in competition. Another result which points to this conclusion is that
50
there was no difference in sugar water recruitment between species in control colonies,
but crazy ants recruited in higher numbers than fire ants in competition colonies (Table
3.1). In contrast to this finding, crazy ants had significantly less mortality than fire ants
in control colonies, but not in competition colonies (Table 3.1, Figure 3.1). This suggests
that fire ants, which experienced heavy mortality in control colonies, were able to inflict a
lot of casualties on crazy ants. Nevertheless, crazy ants had significantly more workers
involved in fights and in the nest box of their competitor than fire ants did (Figure 3.5),
but they did not have significantly more mortality than fire ants. The high aggression but
low mortality for crazy ants is similar to results of aggression assays using 5 workers of
each species (see Chapter 2) and reinforces the findings that aggressive interactions do
not correlate with crazy ant mortality.
Standardizing factor (worker number or biomass) did not significantly affect most
responsible variables, including mortality, resource consumption, and number of ants
fighting (Table 3.1). The one exception was that colonies standardized by biomass,
which had fewer workers than colonies standardized by worker number, had fewer ants
recruiting to both bait types. Additionally, the lack of significant interactions between
species and standardizing factor suggest that a slight crazy ant numerical advantage does
not affect crazy ant recruitment and survival when compared to or competing with fire
ants. In contrast, Morrison (2000) found that S. invicta controlled more baits and
inflicted more mortality on the native fire ant, S. geminata, when colonies were paired by
biomass rather than worker number. Solenopsis geminata workers are much larger than
S. invicta workers (1010 workers per gram compared to 1720 workers per gram),
therefore pairing by biomass gave S. invicta a numerical advantage. However neither
51
biomass nor worker number pairings gave red imported fire ants a numerical advantage
over crazy ants, as crazy ants are much smaller than fire ants, which may explain why fire
ants performed more poorly than crazy ants in both colony unit types. Nevertheless, the
results of the laboratory study, including the finding that standardizing factor was
insignificant, suggest that if a crazy ant population is similar in number to a neighboring
fire ant population, the crazy ants may be able to outcompete fire ants for resources, even
if the fire ants have a greater population biomass than the crazy ants.
The laboratory findings on recruitment times mirror the results of the field
experiment, as crazy ants discovered both bait types more quickly than fire ants in the
field (Figure 3.8). In contrast to the laboratory findings, crazy ant recruitment to both
bait types was much lower than fire ant recruitment (Figure 3.7), and this may be due to
the large difference in field abundance between the species. Relatively small colonies of
crazy ants were introduced to an area dominated by fire ants nests, which likely each
have thousands of workers (Markin et al. 1973). The difference in recruitment
abundance appears to be more important than recruitment time in determining control of
baits, as fire ants controlled significantly more protein baits (which received the highest
recruitment by both species) at the end of 1 hour. Also, in the field, ants of both species
recruited to protein in higher numbers, though no preference for protein was shown in the
laboratory experiment (Figure 3.7). This may be because laboratory colonies were fed a
diet that included generous amounts of protein until 24 hours before the start of the
experiment, whereas field colonies were likely protein-limited, as ants are often more
abundant than prey items (Tobin 1991,1994). Also, laboratory colonies had negligible
brood, whereas field colonies presumably had brood, which have a greater need for
52
protein than adult workers (Holldobler and Wilson 1990). Ant species did not co-occur
at protein baits, suggesting that protein resources, even when available, may be difficult
to control in the field due to high levels of competition, thus increasing a colony's need to
recruit in high numbers.
In a field setting, crazy ants discover baits more quickly but fire ants dominate
baits after 1 hour. Thus, in this context, crazy ants would be considered opportunists,
who discover food quickly and exploit it before other ants arrive, while fire ants would be
considered extirpators that aggressively dominate discovered resources (Wilson 1971).
Both species fit the dominance-discovery tradeoff model in the field setting (Fellers
1987). However, in the laboratory experiment, crazy ants discovered baits more quickly
and controlled more baits than fire ants did at the end of the experiment. This finding
suggests that when crazy ant colonies are evenly matched in worker number or biomass
with fire ants, crazy ants may be both the discoverers and the controllers of baits.
Therefore, crazy ants may have a similar community role as Argentine ants, which have
also shown exploitative competitive ability as well as behavioral dominance in part of
their invasive range (Holway 1999). Nevertheless, ant community roles, including those
of opportunists and extirpators, can be context-dependent (Morrison 1996), and our
findings are an example of this phenomenon, as crazy ant abundance appears to be
important in determining their role as opportunists or extirpators. Our findings are
similar to that of Walters and Mackay (2005), who showed that Argentine ants were able
to overcome the native ant Iridomyrmex 'rufoniger' in laboratory trials only when
Argentine ant colonies were 5 to 10 times larger than Iridomyrmex colonies. Holway and
Case (2001) also showed the importance of colony size on exploitative and interference
53
ability in Argentine ants, as every measure of competitive performance increased with
worker number. As shown in Chapter 2, fire ants are more successful than crazy ants in
individual interactions, yet our findings in the laboratory experiment show that crazy ants
can be more successful at a colony level when colony units are closely matched by
biomass or worker number. Our field results reinforce the importance of relative crazy
ant abundance, as crazy ants recruited to resources in much lower numbers than fire ants
and behaved as opportunists, rather than controlling the baits which they discovered first.
In the field experiment, fire ant recruitment decreased in both control and crazy
ant addition locations (Figure 3.6). Though crazy ant recruitment increased through time
in the crazy ant addition treatment, our data suggest that the decrease in fire ant
recruitment is not due to the introduction of crazy ants. Simply, the addition of crazy ant
colonies did not cause an immediate and effective displacement of fire ants. Fire ant
decline may have been due to environmental factors such as the drought and extreme heat
in the weeks prior to and during the duration of the field experiment, when it rained only
once. In a laboratory and field experiments, fire ant foraging and bait retrieval has been
shown to decrease in increasing temperatures (Drees et al. 2007, Zheng et al. 2007).
Field temperatures often exceeded 35°C in June and July, which may have
contributed to the decrease in fire ant recruitment during the experiment. Additionally,
fire ants have been found to have higher summer mortality (personal communication, M.
Eubanks). Therefore, crazy ants, which have been reported to increase in abundance
during the summer months (personal observation, T. Rasberry), are likely responding to,
rather than driving, a decrease in fire ant recruitment. Reports in the popular press (such
as Blumenthal 2008) that crazy ants rapidly displace fire ants are likely not true. The
54
pattern of a switch from fire ant dominance to crazy ant dominance appears to be correct
but the proposed mechanism is not. One possible explanation for the lack of fire ant
replacement by crazy ants is that crazy ant colonies used in crazy ant addition locations
may not have been large enough to send enough workers to baits to deter fire ants.
However, retesting with larger colonies of crazy ants would be difficult due to the
logistical and ethical considerations that must be a part of site selection.
Crazy ants have been shown to be less effective fighters in individual contests
(see Chapter 2), but they did not differ from fire ants in mortality in competing colonies,
and they recruited to baits more quickly and more abundantly than fire ants. Therefore,
fire ants do not have a universal competitive advantage over crazy ants. The ability of
crazy ants to defend food resources from fire ants may be related to both numerical
abundance as well as habitat. Environmental conditions, including soil moisture and
stream flow, have been shown to affect the spread of Argentine ants (Holway 1998,
Menke et al. 2007), so it is not unlikely that environmental condition could affect the
distribution of crazy ants. Even extremely large supercolonies may not be able spread
out of the woods, as open spaces adjacent to heavily invaded woody areas generally have
low to moderate amounts of crazy ants (personal observartion, K. Horn). Abandoned fire
ant mounds, but no active mounds, can sometimes be found in woods heavily invaded by
crazy ants, suggesting that crazy ants may have displaced fire ants in those areas but we
cannot confirm this pattern from our experiments. Regardless, active fire ant mounds are
not found in wooded areas with high crazy ant abundances, suggesting that at the very
least, crazy ants are preventing re-establishment of fire ants in a formerly suitable habitat.
These clues suggest that in addition to an increased competitive ability with increasing
55
abundances, crazy ants may be competitively superior to fire ants in woody habitats
where they manage to become established, but environmental conditions may prevent
them achieving high population densities and displacing fire ants in open areas. Though
further studies would need to be conducted to confirm this hypothesis, there is an
assortment of features of woody habitats that could be important to crazy ant success,
such as thermal conditions, litter structure, and availability of sugar resources like
honeydew produced by sap-feeding insects. Sugar availability in particular may be
important in determining suitable habitat for crazy ants. Argentine ants, which display
many characteristics similar to crazy ants as previously noted, were found to establish
nests in areas where they had not previously been able to invade following the addition of
sucrose feeding stations (Rowles and Silverman 2009). In the case of crazy ants, woody
habitats may provide the environmental conditions and resource availability necessary to
achieve high densities. A large-scale survey of sites invaded by crazy ants as well as
studies examining interactions between fire ants and crazy ants at invasion fronts would
be useful in experimentally testing this hypothesis.
Clearly, the competitive abilities of ants are not tested before they are introduced
to a new range, yet for new introductions that cannot be eradicated, studies examining
competitive interactions between introduced ants and established ants in the introduced
range may be useful. Such studies could help assess the potential invasive risk associated
with different species. In some scenarios, this could aid in the development of a
management plan, whether it be active control of a high risk species, or a more passive
approach for species which are less competitively dominant and may not require the use
of finite resources reserved for invasive species control. Our findings suggest that colony
56
size and habitat type may be key factors in the crazy ant invasion. Crazy ants can coexist
with fire ants in low abundances in open habitats (as shown with pre-crazy ant addition
field baiting results), yet observations have shown high abundances of crazy ants and an
absence of fire ants in invaded woody areas. Crazy ants are more successful in clashes
with fire ants on a colony- rather than an individual-level, yet relative abundance alone
cannot explain the invasion dynamics, as open areas adjacent to heavily invaded
woodlands are often nearly absent of crazy ants. Further studies examining invasion
front dynamics may be useful in testing this hypothesis, yet our findings are a key
preliminary step that suggests that crazy ant invasions may be restricted to wooded areas
and areas free of fire ants, as they are competitively excluded by fire ants in open areas.
Beyond the implications for Rasberry crazy ant management, this study has
important implications for the field of invasion biology. As the first study to
experimentally examine biotic resistance to a newly introduced species by an established
invasive species, these findings demonstrate the importance of competitive interactions
between invasive species in the introduced range, which have been largely ignored (but
see Belote and Weltzin 2006 and Griffen et al. 2008). The results of this study give
compelling reason for further study of antagonistic interaction between introduced
species, as research in this area may help not only with species-specific applied issues,
but also with our understanding of invasion biology as a whole.
Acknowledgements
I would like to thank Matt Barajas, Juli Carrillo, Marilu Corona, Christopher
Gabler, Johanna Horn, Summer Jones, Matt King, Maria Meza-Lopez, Max Quintanilla,
57
Evan Siemann, and Erica Soltero for their assistance in the lab and in the field. Financial
support was provided by a Wray-Todd Fellowship to KH.
Literature Cited
Belote, R. T. and J. F. Weltzin. 2006. Interactions between two co-dominant, invasive
plants in the understory of a temperature deciduous forest. Biological Invasions
8:1629-1641.
Benjamini, Y. and Y. Hochberg. 1995. Controlling the false discovery rate - a practical
and powerful approach to multiple testing. Journal of the Royal Statistical
Society: Series B 57:289-300.
Blumenthal, R. 16 May 2008. A pest without a name, becoming known to ever more.
New York Times.
Buczkowski, G. and G. Bennett. 2008. Aggressive interactions between the introduced
Argentine ant, Linepithema humile and the native odorous house ant, Tapinoma
sessile. Biological Invasions 10:1001-1011.
Davidson, D. W. 1998. Resource discovery versus resource domination in ants: a
functional mechanism for breaking the trade-off. Ecological Entomology 23:484-
490.
Drees, B. B. M., B. Summerlin, and S. B. Vinson. 2007. Foraging activity and
temperature relationship for the red imported fire ant. Southwestern Entomologist
32:149-155.
Fellers, J. H. 1987. Interference and exploitation in a guild of woodland ants. Ecology
68:1466-1478.
58
Griffen, B. D., T. Guy, and J. C. Buck. 2008. Inhibition between invasives: a newly
introduced predator moderates the impacts of a previously established predator.
Journal of Animal Ecology 77:32-40.
Holway, D. 1998. Effect of Argentine ant invasions on ground-dwelling arthropods in
northern California riparian woodlands. Oecologia 116:252-258.
Holway, D. 1999. Competitive mechanisms underlying the displacement of native ants by
the invasive Argentine ant. Ecology 80:238-251.
Holway, D. and T. Case. 2001. Effects of colony-level variation on competitive ability in
the invasive Argentine ant. Animal Behaviour 61:1181-1192.
Holldobler, B. and E. O. Wilson. 1990. The ants. Springer-Verlag, Berlin, Germany.
Kabashima, J., L. Greenberg, M. Rust, and T. Paine. 2007. Aggressive interactions
between Solenopsis invicta and Linepithema humile (Hymenoptera: Formicidae)
under laboratory conditions. Journal of economic entomology 100:148-154.
Markin, G. P., J. H. Dillier, and H. L. Collins. 1973. Growth and development of colonies
of red imported fire ant, Solenopsis invicta. Annals of the Entomological Society
of America 66:803-808.
Menke, S., R. Fisher, W. Jetz, and D. Holway. 2007. Biotic and abiotic controls of
Argentine ant invasion success at local and landscape scales. Ecology 88:3164-
3173.
Morrison, L. 2000. Mechanisms of interspecific competition among an invasive and two
native fire ants. Oikos 90:238-252.
Morrison, L. W. 1996. Community organization in a recently assembled fauna: The case
of Polynesian ants. Oecologia 107:243-256.
59
Roulston, T., G. Buczkowski, and J. Silverman. 2003. Nestmate discrimination in ants:
effect of bioassay on aggressive behavior. Insectes Sociaux 50:151-159.
Rowles, A. D. and J. Silverman. 2009. Carbohydrate supply limits invasion of natural
communities by Argentine ants. Oecologia 161:161-171.
Tobin, J. E. 1991. A neotropical rainforest canopy, ant community: some ecological
considerations, in C. R. Huxley and D. F. Cutler, editors. Ant-plant interactions.
Oxford University Press, Oxford, UK.
Tobin, J. E. 1994. Ants as primary consumers: diet and abundance in the Formicidae. in
J. H. Hunt and C. A. Nalepa, editors. Nourishment and evolution in insect
societies. Westview Press, Boulder, CO.
Walters, A. and D. Mackay. 2005. Importance of large colony size for successful
invasion by Argentine ants (Hymenoptera: Formicidae): Evidence for biotic
resistance by native ants. Austral Ecology 30:395-406.
Wilson, E. O. 1971. The Insect Societies. Belknap Press, Cambridge, Massachusetts.
Zheng, J. H., R. Q. Mao, and R. J. Zhang. 2007. Comparisons of foraging activities and
competitive interactions between the red imported fire ant (Hymenoptera:
Formicidae) and two native ants under high soil-surface temperatures.
Sociobiology 50:1165-1175.
60
Table 3.1. Tests for each response variable for each comparison. P< 0.05 are in bold.
Factor = standardizing factor (biorrtass or worker number).
Mortality Controls
Competition (RCA-RIFA)
ANOVA
ANOVA t-test
Species Factor
S*F Factor
mean=0
Fi,26=5.66 F,,26=1.91 F,.26=1.44 FU4=2.10 t I5=-1.49
p=0.0250 p=0.1792 p=0.2413 p=0.1696 p=0.1581
Mealworm consumed Controls
Competition
ANOVA
Wilcoxon
Species Factor
S*F Factor
FU7=2.37 F1,27=0.0001
F127=0.56 Zi=1.00
p=0.1357 p=0.9912 p=0.4600 p=0.3184
Sugar water consumed Controls
Competition
ANOVA
ANOVA
Species Factor
S*F Factor
Fi.aF5.72 FU6=0.52 Fi.26=0.02 Fi.i4=0.26
p=0.0243 p=0.4777 p=0.9014 p=0.6175
Mealworm total recruitment Controls
Competition (RCA-RIFA)
ANOVA
ANOVA t-test
Species Factor
S*F Factor
mean=0
Fi 27"~6.81 Fi'27=4.72 Fi.27=1.50 Fi,i4=0.H t,'5=3.35
p=0.0146 p=0.0387 p=0.2317 p=0.7409 p=0.0044
Sugar water total recruitment Controls
Competition (RCA-RIFA)
ANOVA
ANOVA t-test
Species Factor
S*F Factor
mean=0
Fi,27=0.29 FU7=0.002 Fi.27=0.03 FU4=0.04 t,5=4.27
p=0.2917 p=0.0024 p=0.0299 p=0.8484 p=0.0007
Mealworm recruitment time
Controls Competition (RCA-RIFA)
Wilcoxon
Wilcoxon
Wilcoxon
Species
Factor
mean=0
Z,=3.63
Z,=0.37
Z15=-51.00
p=0.0003 p=0.7120
p=0.0019 Sugar water recruitment time
Controls
Competition (RCA-RIFA)
Wilcoxon
Wilcoxon Wilcoxon
Species
Factor
mean=0
Z,=2.68
Zi=-0.37 Z,5=-36.00
p=0.0074 p=0.7101 p=0.0024
Total ants fighting Competition (RCA-RIFA) ANOVA
t-test Factor
mean=0 FU 4=1.08 t15=5.37
p=0.3159 p<0.0001
Total ants in other nest box Competition (RCA-RIFA) Wilcoxon
Wilcoxon
Factor
mean=0 ZpM.19 Z7= 18.00
p=0.2330
p=0.0078
Final bait control
Competition Binomial Species p=0.0327
Table 3.2. Significant correlations between response variables. All correlations with
discovery and recruitment times were non-normal, and therefore the correlation value is
Spearman's p. The Benjamini and Hochberg FDR method (Benjamini and Hochberg
1995) was used to correct for multiple comparisons. MW = mealworm and SW = sugar
water.
Variable 1 SW recruitment
SW recruitment time SW recruitment time MW recruitment time MW recruitment time MW recruitment time
Variable 2 MW recruitment MW recruitment SW recruitment MW recruitment
Mortality SW recruitment time
Correlation 0.4977 -0.6936 -0.5298 -0.4964 0.4008 0.8359
p-value 0.0053
<0.0001 0.0031 0.0116 0.0471
<0.0001
62
Figure 3.L Number of dead ants at the end of the colony-level competition experiment.
Fire ants (RIFA) experienced higher mortality than crazy ants (RCA) in control colonies
(p = 0.0250), but not in competition colonies (p = 0.1696). Because the standardizing
factor affected the total number of ants in a colony unit, the number of dead ants for each
species and treatment is divided into worker number colony units and worker biomass
colony units. Error bars are +1 SE.
* - » c CO
" O CO CD
_ Q
E
180
160
140
120
100
80 H
60
40
20 H
0
Control
RCA
Competition
i i Worker Biomass i i Worker Number
I
RIFA RCA
Ant Species RIFA
63
Figure 3.2* Average sugar water consumption by species, adjusted for average
desiccation. Fire ants (RIFA) consumed more sugar water than did crazy ants (RCA, p =
0.0243). Average crazy ant sugar water consumption was negative due to the method of
calculation (described in detail in the Methods). Error bars are ±1 SE.
0.16
P> o„14
0.12 H
0.10
0 08
0.06
0.04
0.02
o Q.
E CO c o o (D CO
(0 0.00 o> 5 =0.02
CO -0.04
RCA RIFA
Species
64
Figure 3.3. Recruitment to mealworms and sugar water. Crazy ants (RCA) recruited to
mealworms, but not sugar water, in higher number than fire ants (RIFA) in control
colonies (p = 0.0146 and p = 0.2917, respectively). Crazy ants recruited to both baits in
higher abundances than fire ants in competition colonies (p = 0.0044 for mealworm, p =
0.0007 for sugar water). Error bars are +1 SE.
(/)
CO
60
50
40
o »- 30 0
E =3 20
10
Control colonies
I Ix
Competition
1 Mealworm l I Sugar water
J , i
RCA RIFA RCA RIFA
65
Figure 3.4. Mealworm and sugar water recruitment times. Crazy ants (RCA) recruited
to mealworms and sugar water more quickly than fire ants (RIFA) in control colonies (p
= 0.0003 for mealworm, p = 0.0074 for sugar water), as well as in competition colonies
(p = 0.0019 for mealworm, p = 0.0024 for sugar water). Error bars are +1 SE.
7000
j D 6000 13 C
"C 5000
O C 4000
P " £ 3000
E b 2000
o 0 1000
a:
Control colonies
i i Mealworm I I Sugar water
X,
X
Competition
RCA RIFA RCA RIFA
Figure 3.5. Ants fighting and ants in the other nest box. Crazy ants had more ants
fighting (p < 0.0001) and more ants in their competitors' nest box (p = 0.0078) than fire
ants did.
100
80 H
-i—»
£ 60 O
<D
"§ 40 13
20
i i Crazy ants i i Fire ants
fighting in other nest box
Figure 3.6. Changes in ant recruitment to field baits through time. Sampling began on
June 24, five days after treatment establishment, and ended on July 14, twenty-five days
after treatment establishment.
- •— RCA addition • o - control
June 24 June 29 July 3 July 8 July 14
Sampling Date
68
Figure 3.7. Average ant recruitment to field baits by treatment and bait type from June
24 to July 14. Recruitment was significantly higher to protein than to sugar for crazy ants
(p = 0.0126) and for fire ants (p < 0.0001). On average, crazy ant recruitment was higher
in crazy ant addition locations (p = 0.0610), but treatment had no effect on fire ant
recruitment (p = 0.8718). Error bars indicate + 1 SE.
100 i
80
+-»
£ 60
o <D
"? 40 13
20 X
_r±i
I
n Crazy ants 3 Fire ants
protein sugar
Bait Type
x
control addition
Treatment
69
Figure 3.8. Discovery time of field baits by bait type and species. Sugar baits were
discovered significantly more quickly than protein baits (p = 0.0140), and crazy ants
discovered baits of both type more quickly than fire ants (p = 0.0092). Error bars indicate
+ 1SE.
0)
C
a> E
> O o b
10 -,
8H
4H
2H
protein sugar
Bait Type crazy ants fire ants
Species
70